The following commentary is part of First Monday's Special Issue #2: Open Source.
Although the phenomenon of Free/Libre Open Source Software (FLOSS) development is now several decades old, empirical research in this area is still developing. The papers presented in this special issue suggest how much we have learned, but as well how much more remains to be learnt. Looking over these papers suggests two areas that are particularly important for future empirical research on FLOSS development.
First, much research has viewed the development process as a black box and studied only inputs, such as developer motivations, and outputs, such as code quality. It is time to open the black box to examine the development process in detail. There have been a number of illuminating case studies of these processes, but few studies compare across more than 1 or 2 projects to provide a generalizable picture. Since FLOSS development is generally undertaken by teams of developers, empirical studies of the development process need to be informed by a deep theoretical understanding of how members of teams (especially distributed teams) work together in general. As well, studies need to take into account the particular characteristics of FLOSS teams: their geographic distribution and technological support, frequent reliance on volunteer members, and the existance of a small high-contributing core of members and a larger lower-contributing periphery. Questions for continuing research include:
- how the development work is coordinated in FLOSS teams, e.g., to identify FLOSS-specific coordination mechanisms used to manage dependencies in a distributed group;
- how shared understandings (of the system, users, other developers) support development and how these understanding develop among the core and peripheral members;
- the role of leadership and its development, especially given that many FLOSS teams incorporate a large number of volunteers;and
- interaction between voluntary and paid employees as the involvement of private sector in the Open Source projects increases.
As well, future studies should analyze development teams over time to provide a dynamic view of the practices of FLOSS projects. Such studies can link empirical work examining patterns of change within a code base, often called ‘software evolution’, to patterns of change within the development team and the technologies that support its collaboration.
A second issue that empirical studies of FLOSS development need to address is the building of a cumulative science. There are a growing number of studies, but in many cases, these talk past rather than building on each other. In order for our work to cumulate, research is needed on:
- development and validation of agreed-on measures for important aspects of the FLOSS development. For example, it is currently hard to assess if a project is successful, which makes it more difficult to identify what practices projects might adopt to improve their chances of success or to pursue the much needed study of failed projects.
- development of common data sets. Being able to carry out different studies on the same sample of projects or developers will enable more precise comparison of results.
- characterization of different kinds of FLOSS projects. FLOSS and its development is generally (often implicitly) defined as projects that use a free or open source license, but there are a wide range of system types, developer goals, software development processes, management styles and group structures, all of which are compatible with releasing software under an OSI approved license. Lumping together a diversity of projects is certain to dilute any conclusions that might be drawn about them.
The increase in empirical studies reflected in the papers in this special issue mirrors the increasing popularity of FLOSS itself. We hope that we as researchers can do as good a job in developing our knowledge of how FLOSS development works as the FLOSS developers have done developing software.
Copyright ©2005, First Monday
Copyright ©2005, by Kevin Crowston
Future research on FLOSS development
First Monday, Special Issue #2: Open Source (October 2005),